We are proud to share that One Place has been awarded research and development funding through the Estonian Applied Research Programme (Rakendusuuringu ja tootearenduse toetus), administered by Enterprise Estonia (EIS). The grant supports our project titled "Combining computer vision and graph neural networks for accurate, scalable real-estate valuation" and marks an important milestone in our journey to transform how people search for and evaluate property across Europe.
The project was evaluated by independent experts under the programme's fourth call for small projects, exceeding the required threshold across all selection criteria. This recognition validates both the scientific rigour and commercial viability of the approach we have been developing.
At its core, the project aims to develop and validate an AI-based prototype that fundamentally changes how real estate data is discovered and understood. Today, property information is scattered across hundreds of portals, each with its own format, its own quirks, and its own blind spots. A buyer searching for a home in Tallinn, Lisbon, or Amsterdam is forced to navigate a maze of incompatible listings, often with little more than a few photos and a price tag to guide their decision.
Our approach tackles this fragmentation head-on. We are building custom computer-vision models that can look at a property photograph and extract meaningful information: the condition of the interior, architectural style, the presence of specific features like a renovated kitchen or original hardwood floors. Paired with graph neural networks, these visual signals are combined with structured data and spatial context to produce accurate, explainable property valuations.
The expected outcome is the technological validation of these models at TRL 3 to 5, proving that automated feature extraction and intelligent price estimation can work at scale and at a fraction of current market costs. If successful, this technology will allow anyone, whether a first-time buyer or a seasoned professional, to understand the true value of a property with the kind of depth and nuance that was previously reserved for experienced local agents.
The project runs from December 2025 through March 2027. Over this period, our research team will train and validate the core models, integrate them into the One Place platform, and measure their performance against real market data. Every step brings us closer to our mission: unifying the European real estate market into one intelligent, transparent search experience.
We are grateful to Enterprise Estonia for their confidence in our vision and to the independent experts who evaluated the proposal.
